Nano Futures

PAPER

Femtosecond photonic viral inactivation probed using solid-statenanoporesTo cite this article: Mina Nazari et al 2018 Nano Futures 2 045005

 

View the article online for updates and enhancements.

This content was downloaded from IP address 155.33.129.30 on 14/11/2018 at 20:04

Nano Futures 2 (2018) 045005 https://doi.org/10.1088/2399-1984/aadf9d

PAPER

Femtosecond photonic viral inactivation probed using solid-statenanopores

MinaNazari1,2, Xiaoqing Li3,MohammadAminAlibakhshi4, Haojie Yang5, Kathleen Souza6,ChristopherGillespie6,11, SuryaramGummuluru7,MiKHong8, BjörnMReinhard2,9, Kirill SKorolev8,10,LawrenceDZiegler2,9, Qing Zhao3,MeniWanunu4,12 and Shyamsunder Erramilli2,8,12

1 Department of Electrical andComputer Engineering, BostonUniversity, Boston,MA02115,United States of America2 Photonics Center, BostonUniversity, Boston,MA02115,United States of America3 School of Physics, PekingUniversity, Beijing, People’s Republic of China4 Department of Physics, NortheasternUniversity, Boston,MA02115,United States of America5 Department ofMechanics, Southeast University, Nanjing, People’s Republic of China6 Next Generation Bioprocessing,MilliporeSigma, Bedford,MA01730,United States of America7 Department ofMicrobiology, BostonUniversity School ofMedicine, Boston,MA02118,United States of America8 Department of Physics, BostonUniversity, Boston,MA02115,United States of America9 Department of Chemistry, BostonUniversity, Boston,MA02115,United States of America10 Department of Bioinformatics Program, BostonUniversity, Boston,MA02115,United States of America11 Current address of CGillespie: Immunogen, 830Winter StWalthamMA,United States of America.12 Authors towhomany correspondence should be addressed.

E-mail: wanunu@neu.edu and shyam@bu.edu

Keywords:nanopore, virus, plasmon, femtosecond, laser

Supplementarymaterial for this article is available online

AbstractWe report on detection of virus inactivation using femtosecond laser radiation bymeasuring theconductance of a solid state nanopore designed for detecting single particles. Conventionalmethodsof assaying for viral inactivation based on plaque forming assays require 24–48 h for bacterial growth.Nanopore conductancemeasurements provide information onmorphological changes at a singlevirion level.We show that analysis of a time series of nanopore conductance can quantify the detectionof inactivation, requiring only a fewminutes from collection to analysis.Morphological changes wereverified by dynamic light scattering. Statistical analysismaximizing the information entropy providesameasure of the log reduction value. This work provides a rapidmethod for assaying viral inactivationwith femtosecond lasers using solid-state nanopores.

1. Introduction

Existing and emerging viruses are amajor threat to human and veterinary public health. The need for safe andreliable inactivation or removal of viruses is universal in antiviral therapies, pharmaceuticals, and viral vaccinedevelopment. Conventional pharmaceutical pathogen inactivationmethods are quite effective, but they involvesubstantial collateral damage and have undesirable side-effects [1–3]. Chemical-free viral inactivationmethodssuch as ultraviolet and gamma-irradiation have been used tominimize some of the side-effects. Nevertheless,thesemethods adversely affect thermolabile compounds and denature biomolecules of interest in themediumcontaining the virus. Ultrashort pulsed lasers (UPL) provide newopportunities for chemical-free pathogendisinfection in solution. Photonicmethods have the potential to provide an attractive alternative to existingbiocides and ionizing radiation techniques [4–8]. Photonic inactivation has been successfully achievedwithfocused femtosecond (fs) laser pulses for exposure times of�1 h on sample volumes typically of�2 ml [5–10].Although the ultrafast laser inactivationmethod for viral inactivation is fairly well established, a systematicunderstanding of the inactivationmechanism is currently lacking. There is a need for rapidmethods for assayingviral inactivation in order to carry outmulti-dimensional studies of the ultrafast laser parameters such as

RECEIVED

21August 2018

ACCEPTED FOR PUBLICATION

7 September 2018

PUBLISHED

24 September 2018

© 2018 IOPPublishing Ltd

intensity, pulse duration and center wavelength that can contribute to the design and optimization ofinactivation protocols.

Viral inactivation is a complicated process and its outcomehighly depends on the specific treatmentmethod. Awide range of biological assays coulddetect andquantify intact viruses in an ensemblemannerwhich is extremelylaborious, time-consuming,with low sensitity [11]. A different approach is to explore viruses at the single virionlevel.Different opticalmethods have been developed to characterize single viral particles; [12–14]but still there is aneed for a technique that is fast, sensitive anduses small sample volumes. Although imaging techniques, such asAFMandTEM, are capable of characterizing viruseswith high sensitivity, resultswill be inevitably affected by thetedious and costly sample preparation steps.Nanometer-sized pores in amembrane offer the capability ofelectrically detectingmolecules in a label-freemanner at single-molecule level in a volume as small as a fewmicroliters anddetection times as short as a few seconds. Passage ofmolecules andparticles througha nanoporecauses transient disruption in the ionic current through it, fromwhich the size, concentration, anddistributionofanalytes can bededuced [15, 16]. The electrical signal characteristics in a given analyte sample strongly dependonthe analyte passing through the pore, thepore geometry, and the experimental conditions such as pH, ionicstrength, applied voltage, temperature, etc. This single-particle electrical sensorhas beenused for quantifying theconformational properties of proteins [17–23], understandingDNA transport [24–26] anddetecting smallmolecules [27, 28], amongmany other applications. Previous studies have reported the ability of nanopore sensorsto detect spherical and icosahedral viruses [29], virus capsids [30], themasses and zeta potentials of viruses [31, 32],and to explore the translocation of stiff, rod-shaped viruses [33].

This ability of nanopores to detect the translocationof nanometer size particlesmotivatedus to study the effectof anoptical viral therapyon the single virus level, crucial for the preparation of very safe biotherapeutics. In thispaper, high incident fs laser pulse intensities of>100 GW cm−2, which aremore than 105 times greater thanusedinprevious studies, have beenused to inactivate viruses, leading to 4-log reduction in viral activity in 1 minirradiation of∼2ml sample volume. This result showsnearlymore than twoorders ofmagnitude improvement intreatment time compared to conventional pulsed laser viral inactivationmethods [34]. Furthermore,wedemonstrate the capability of the nanopore technique to precisely characterize individual viruses, explore howvitalviral function is affected by treatment, and quantify the effectiveness of this label-free viral inactivation technique.In light of these points, we investigate the effects of fs laser on inactivatedΦX174 bacteriophage,whichhas thefirstsequencedDNA-based phage genomewidely used standard for viral clearance, aswell as a surrogate for enterichuman viruses [35]. Changes in thephysical properties of treated virus samples aremonitoredby electricallycounting viruses in a small sample volume.A statistics-basedmethodhas been developed tomonitor the reductionvalue of viruses using sequential nanoporemeasurements, and comparedwith a plaque forming assay.Moreover,the effect of inactivationonviruses in an ensemblemanner and at the single-virus level resulting fromdynamiclight scattering (DLS)measurements andnanopores, respectively, are compared.

2.Methods

2.1. Femtosecond laser irradiationFemtosecond laser pulses from a Legend EliteDuo (Coherent Inc.)ATi-sapphire regenerative amplifier systemwas used as the excitation source in this study. This laser produces a continuous train of 35 fs pulses at arepetition rate of 1 kHz. The output of the second harmonic generated by a~ 1mm thick BBO crystal with awavelength centered at∼400 nmwas used to irradiate the virus samples, with a combined pulse energy of thesecond harmonic and fundamental of about 2.5mJ. Figure 1(a) depicts the experimental setup. The laser beamwith spot size∼1 cm2was incident upon a typically 1 cmquartz cuvette containing 2 ml of virus sample while astirrer was used to homogenize the virus’s interactionwith the laser beam. ForΦX174 sample with 1012 pfu ml–1

concentration, the laser treatment ismade by exposing 250 μl of viruses in 2 mmcuvette. The typical sampleexposure timewas 15 minutes. All experiments were carried out at 22 °C, and all samples were immediatelystored at 4 °Cafter irradiation. Experiments were carried out in triplicate.

2.2. Virus sample preparationΦX174 samples with 2×1012 plaque forming unit (pfu) perml concentration (Promega TiterMaxΦX174Bacteriophage) in 0.05 M sodium tetraborate were stored at−80 °C. Before the experiment, samples werethawed to room temperature, aliquoted, and kept at 4 °C. For diluted samples,ΦX174 spiked feed solutionswereprepared by serial dilution in Sorenson’s buffer to thefinal concentration of approximately 106 pfu ml–1.

2.3. Infectious plaque assayTo count theΦX174 in the solution, samples were dilutedwith Sorensen’s buffer dilution blanks, to bring plaquetowithin a statistically valid range of 30–300 plaques per plate. Samples were assayed in triplicate by adding

2

Nano Futures 2 (2018) 045005 MNazari et al

0.1 ml of diluted sample and 0.1 ml of host cell suspension to a test tube containing 3 ml ofmolten (46 °C–48 °C)ΦX174 overlay agar consist of 10 g of tryptone peptone (Difco), 8.5 g of agar (Difco) and 5 g ofNaCl per liter ofreagentwater. Then the solution containing host cell and bacteriophagewas vortexed and transferred to theΦX174 bottomplate agar (2.5 g ofNaCl, 2.5 g of KCl, 10 g of tryptone peptone fromDifco, 10 g of agar fromDifco, and 1 ml of 1 MCaCl2 per liter of reagent gradewater) and incubated overnight at 37 °C. Then plaqueswere counted and the corresponding bacteriophage concentrations were reported as pfu mL–1. Sorensen’sphosphate buffer (pH7.3),ΦX174 bottomplate agar, andΦX174 overlay agar were purchased fromNortheastLaboratory Services (Winslow,ME).

2.4. Nanopore device fabrication andmeasurementOur electrical detection system is composed of a nanopore formed in a 50 nm-thick insulating silicon nitride(SiN)membrane. The SiN is deposited on a silicon substrate with 2 micron silicon dioxide previously grown on,which is chemically etched by potassiumhydroxide (KOH) to obtain a freestandingmembrane. The electronbeamof a JEOL 2010F transmission electronmicroscope (TEM)was finely focused on themembrane in order tomake a porewith controlled size [24]. The choice of 38 nmnominal diameter of the nanopore was based on theknown radius of the virus derived from its x-ray structure. Several slightly smaller and larger nanopores weretested to obtain the optimal pore size. Amore detailed systematic studywas not attempted. The as-fabricatednanopore chip is thenmounted in a fluoropolymer cell that allows electricalmeasurement of ionic currentthrough the nanopore. The cell isfilledwith 0.1 MKCl solution (16.1 mS cm−1 conductivity), buffered to pH 7using 10 mMtris. The silver–silver chloride (Ag/AgCl) electrodes are inserted in both cis- and trans-chambers,and aDC voltage is applied toflow current and drive chargedmolecules through the pore.

2.5.Data acquisition and analysisAChimeraVC100 (Chimera Instruments LLC)was used for recording the ion current through the nanopore.Datawas digitized at 4.17 MS s−1, and saved to the computer at a 1 MHz bandwidth. Prior to analysis, recordingswere further filtered using a 100 kHz digital low-passfilter. To verify the pore’s stability, before the introductionof a virus sample to the nanopore, several seconds of current were collected to ensure that no event is detectedand the baseline current is stable. Three key independent parameters are extracted from the nanopore data: thedwell time of viruses at the pore, td, the fractional current blockade, FI, and the inter-event waiting time, δt, fromwhich virus capture rates can be extracted.

Figure 1. (a) Schematic of experimental setup for photonic viral inactivation. The following abbreviation are used: fs: femtosecond,QC: quartz cuvette, BBO: Beta bariumborate. (b)Log reduction value (LRV)measured forΦX174 virus with 106 pfu ml–1 (patterned)and six-orders ofmagnitude increased concentration (solid) afterUSP laser irradiation for different exposure time.

3

Nano Futures 2 (2018) 045005 MNazari et al

2.6.Dynamic light scattering (DLS)In order to obtain information on the size distribution of viruses in an ensemblemanner, we performedDLSmeasurements using the ZetasizerNano S90 fromMalvernCorp. ThisDLSmeasurement is based on theBrownianmotion of spherical particles; using the Stokes–Einstein relation to determine the particle size basedonmeasured diffusion constant of particles [36]. For the sizemeasurement,ΦX174 viruswith 1012 pfu ml–1

concentration andΦX174VirionDNA (NewEngland Biolabs)with 1000 μg ml–1 concentration, are diluted by8-fold and 30-fold respectively in 10 mMTris. An aliquot of about 80 μl of the diluted samples is transferred tothe cuvette forDLS analysis with themeasurements performed at 23 °C.

3. Results and discussion

Obtaining extremely high levels of viral clearance is a substantial step in the purification of protein-basedtherapeutics. The presence of even a single virus in the final drug product could be harmful to the consumer’shealth. To prevent this, implementation of an effective viral inactivation strategy is crucial. USP viralinactivationwith greater than 4-log reduction in viral infectivity would enable a new chemical-free pathogenclearance technology [34, 37, 38]. Thefirst objective of this study is to extend thework done onUSP inactivationof viruses, using a regeneratively amplified laser system.We expedite the reportedUSP photonic inactivation(>1 h) [34] to 1 min. In this study, 35 femtosecond pulsed laser irradiationworking at∼400 nm is used toirradiate 2 ml ofΦX174 bacteriophage for different irradiation times. Inactivation of viruses ismeasured by aviral infectivity assay. Original sample concentrations were calculated bymultiplying the plate count by thedilution factor, reported as pfu ml–1 [39].

Thefinal results for the viral inactivation experiments is reported as the log reduction value (LRV)whichprovides a directmeasure of viral inactivation. The LRVwas calculated according to:

=⎛⎝⎜

⎞⎠⎟ ( )C

CLRV log . 1U

T10

HereCU is the concentration of the untreated sample andCT the concentration of the treated samples exposed tothe laser irradiation as described. Control samples consist of a sample with no laser exposure whichwas keptrefrigerated during the experiment and another samplewhich experienced the same pipetting and stirringcondition as treated sample but without the laser exposure. These control samples never differed significantlyandwere taken to check for loss of titer in the treated suspension. As demonstrated in figure 1(b)more than 104

reduction values ofΦX174with 106 pfu ml–1 concentration is achieved by irradiating 2 ml of virus suspensionwith 2.5 W (average power)~ 400 nm femtosecond laser pulses for different treatment times ranging from1 to15 min. The relatively large laser beamdiameter of the regeneratively amplified laser beam results in fast viraltreatmentwhich can overcome the need for long irradiation times, and remove constraints on the correspondingpractical implementation.

The next goal is to preciselymonitor changes occurred to the treated viruses on the single virus level. A 20 μlaliquot ofΦX174with∼1012 pfu ml–1 is treated for 15 min exposurewith the same laser setup. For this lowvolume of sample, we used amicro quartz cuvette with no stirring. Again, the strong reduction in viral infectivity(LRV>3)was achieved for six-orders ofmagnitude increased virus concentration (figure 1(b)).

As shown infigure 2, theΦX174 viruses are being voltage-driven through a∼38 nmporesmade of SiN. ATEM image of the pore is shown as an inset. The pore size is intentionally chosen close to the virus size to slowdown the translocations and allow for accuratemeasurement of the events. Viruses are electro-osmoticallydriven through the nanopores upon application of a negative bias to the trans chamber. The application ofvoltage results in a steady-state countercurrent of K+ andCl− ions across the pore, which produces a stablebaseline open pore current, I .0 When 0.5 μl of virus sample is added to 50 μl of buffer in the cis chamber, passageof individual viruses through the pore reduces the ionic current, which results in a spike in themeasured current.This volumewas sufficient to generate>2×103 events in 10 s for good statistics in the untreated sample. Thespike contains information about single virions, which can be extracted using statistical analysismethods. Thecurrent ismeasured by aChimera VC100 amplifier which streams 1MHz bandwidth data to a computer at asampling rate of 4.17 MHz, followed by application of a 100 kHz low-pass filter in software to reduce the highfrequency noise dominated by the chip capacitance. The choice of the 100 kHz low-pass filter reflects acompromise that enables detection of events with a sufficient sampling time-resolution of one order ofmagnitude smaller than the virus translocation times, and the loss of higher frequency information.

Figure 3(a) shows continuous current traces obtainedwhenuntreatedΦX174 viruseswere added to the cischamber at 40 mVapplied voltage. Each spike corresponds to transport of a single virus through the∼38 nmpore.The electric signal represents twopredominant current levels, thehigher one corresponds to the openpore current

4

Nano Futures 2 (2018) 045005 MNazari et al

when viruses donot translocate through thepore, I ,0 and the lower corresponds to the virus-occupied level. Basedon themeasured open pore currentwhich is 1.13 nA, thepore conductancewas calculated asG=28.2 nS.Theinset shows amagnified viewof a randomly selected event corresponding to individual intactΦX174. Thetranslocation current,ΔI, which is thedifference between the baseline current and thepulseminimum,dependsonpore geometry and virus size. Basedon the knownnanopore diameter of~ 38 nm, thenanopore lengthwasestimated from theopen pore conductance to beheff~ 35 nm.This value, thinner than the overall initial 50 nmmembrane thickness,h, was still thicker than the expected thickness ofheff=h/3previously found for the verynarrowpores [40]. As shown infigure 3(d), themean fractional current blockade for untreatedΦX174virus ismeasured to be 27%.This number is consistentwith the 29%blockade predicted by the following theoreticalequation, derived in supplementalmaterial is available online at stacks.iop.org/NANOF/2/045005/mmedia,

Figure 2. Schematic of the nanopore setup for virus detection. Pores are fabricated in freestanding SiNmembranes and an externalbias is applied across themembrane to driveΦX174 viruses, potassium ions (yellow dots) and chloride (red dots) ions through thepore. Inset: TEM image of a fabricated nanopore. (scale bar: 20 nm).

Figure 3.Comparison of pre-treatment and post-treatment ofΦX174 virus with a pulsed (35 fs) 800 and 400 nm laserwith an averagepower of 2.5 W for 15 min. (a), (b)Continuous ionic current traces collected using a 40 mVapplied voltage, low-pass filtered to100 kHz. Inset: zoomed in portion of selected representative events. (c) Scatter plot of fractional currentblockade, FI (%),at 40 mV(0.1 MKCl,20 °C) vs dwell time, td. (d)Histograms of fractional current blockade at 40 mV showing decrease after laser treatment.(e)Histograms of dwell time at 40 mVvoltage demonstrates faster translocation of treated viruses.

5

Nano Futures 2 (2018) 045005 MNazari et al

for a 30 nmdiameter spherical virus:

=áD ñ

=- ( )F

I

I

R R

Ra. 2I

o

b o

b

Inwhich = +s sp

R 4o d

h

d

1

p

eff

p2 is the open pore resistance andRb is the nanopore resistance in the blocked state:

s sp sp

= +-

+- -

-

⎛

⎝⎜⎜

⎞

⎠⎟⎟ ( ) ( )R

d

h d

d d d

d

d dh d b

14

4tan . 2b

p

eff

p p p

eff2 2 2

1

2 2

Here d is the virus diameter, dp is the pore diameter, and s is the salt conductivity which for the 100 mMKClbuffer wasmeasured as 16.7 mS cm−1 using the conductivitymeter.

The agreement between theoretical and experimental translocation ratio verifies that the untreated virusesmaintain their shape integrity during transport through the pore. Additionally, a 1Ddrift–diffusionmodel canbe used to describe the dwell time distribution of virus translocation through the nanopores. Fitting theprobability density function of dwell time distribution (td)with the following equation yields two importantparameters: diffusion constant of viruses inside the nanopore,Dpore, and their drift velocity, vd

p=

- -⎛⎝⎜

⎞⎠⎟( )

( )( )P t

h

D t

h v t

t D4exp

4. 3

eff

pore d

eff d d

d pore3

2

Application of this equation, which is based on the assumption of barrier-free transport [41], for theuntreated viruses (figure 3(e)) yields m= -D 0.5 m spore

2 1 and < - -v 10 m s .d4 1

To gain insight into the virus transport kinetics through this nanopore, we compare the in-pore diffusion

coefficient of viruses with a bulk diffusion coefficient, using Stokes−Einstein equation =ph

D k T

d3B where kB is the

Boltzmann constant,T the absolute temperature, η the viscosity of the solution, and d the hydrodynamicdiameter of the virion. UsingDLSwemeasured average diameters of∼30 nmat 23 °C, and accordingly,Dbulk

was calculated as m -14 m s .2 1 The result shows a small reduction inDporewhich can be a consequence of virus-pore hydrodynamic interaction, as previously observedwith protein transport through smaller pores [42].

Upon successful detection of untreated viruses using nanopores, we probed the effect of laser treatment onthe virus sample. A time trace for translocation of laser-treatedΦX174with LRV>3 (figure 1(b)) is shown infigure 3(b). There is a drastic change in the time trace of laser irradiated viruses as compared to untreatedsamples whichwill be explored inmore detail. The averaged baseline conductance remained constant towithin1%over the data collection timescale, indicating that the nanopores did not expand andwere not blocked bydebris or other particles.

To explore the effect of laser treatment, the scatter plot of fractional current blockade versus dwell time oftreated and untreated viruses at 40 mVvoltage is shown infigure 3(c). Two clear groupings of events are visiblynoted, corresponding to the treated and untreated viruses which can be visually distinguished by drawing a lineas shown infigure 3(c). A histogramof fractional current blockades and dwell time distributions for both treatedand untreated samples at 40 mV alongwith generalized extreme value distribution fits to the distributions areshown infigures 3(d)–(e) respectively. The untreated sample is centered at FI=25.04+/−3.1%withlog10(td)=2.05+/−0.36 (tdmeasured inμs) and the treated sample is centered at FI=10.4+/−2.7%withlog10(td)=1.3+/−0.12 . These two sets of independent parameters (FI and td) clearly show the effect of lasertreatment on viruses. The FI is related to the size of the particle and hence a strong decrease in the fractionalblockade yields the first important information on the effect of fs laser treatment, i.e., the global appearance ofthe non-envelopedΦX174 virus does not remains intact after treatmentwhich is in agreement with previouslyestablished data [8].Measurements of the fractional blockadewith slightly larger diameter nanopores(dp∼39–40 nm) gave results similar to those reported infigure 3whereas a smaller diameter nanoporeproduced events with longer dwell times. A systematic study of the pore size dependence on treated anduntreated samples was not attempted, given the limited scope of the present study.

Using the capability of our label-free resistive pulse technique, one canmonitor the inactivation of viruses bylooking for the presence of intact viruses in the solutionwith high sensitivity. One simplemodel is to consider anellipse in the scatter plot of untreated events centered at themean value of fractional current in y direction andthemean of log10(td) in x direction. The semi-minor and semi-major axis radius of this ellipse is equal to threetimes the standard deviation of the FI and log10(td) of untreated events, respectively (figure 4(a)). The standarddeviations are obtained based on theGaussian fits of the corresponding data. Counting the number of thetreated (red) data points in the black ellipse gives us a rough estimate of the number of intact viruses in thetreated sample, n. The number of points in untreated and treated scatter plot is 2723 and 2564, respectively. The

6

Nano Futures 2 (2018) 045005 MNazari et al

ratio of n to the total number of red data points of the treated sample,N, gives us an estimation of the survivalfraction (r)

= ( )rf

f

n

N, 4I

T

where fI and fT are the capture probabilities of intact and treated virus. Assuming these probabilities to be equalin this case, the survival fraction is approximately 0.02. This quantity can provide a simplemethod for rapidlyestimating the efficiency of the inactivation technique.

Amore precise way to estimate rusing nanopore data is to determine an upper bound for the number ofsurvived viruses after treatment. Tofind the upper boundwe developed a statistical formula that employs aprobability distribution function derived from figure 3(d). Thefirst step is to estimate the probability offinding atranslocation event of the treated sample with a distinguishing feature xT inside the untreated sample regionwithxU. In a simplest case xT and xU can be any feature that can separate the two populations; in our case it can befractional current blockade, dwell time or any other combination of these two. For example if we consider thisone dimensional scalar as the fractional current blockade, in our case there is a big separation between themeanof untreated events, x ,U and treated ones, x ,T so that <x x .T U Some of the viruses in the treated sample couldbe intact.We can quantify this possibility by considering the probability ( )P xI T that an untreated virus has avalue of the feature as low as xT, which is the cumulative distribution function for the untreated sample:

=( ) ( )P x CDF x .I T U T Thus, for the treated data points that are far away from the untreated events, the resultwould be zerowhile for the ones which has overlap, the result ismore than zero. By assuming all untreatedviruses are intact, we can consider a threshold, ε, on the probability PI, and using a heaviside step function, θ, wedecide if an event can be counted as an intact virus (if e( )P xI T ) or not (if e<( )P xI T ). By applying thisthreshold on all events, we can define a function ρε as equation 5which is a sumon all treated data pointsnormalized by the number of treated events,N.

Figure 4. (a) Scatter plot of fractional blockade of untreated (green) and treated (red) viruses versus event duration, under a drivingvoltage of 40 mV (b) upper bound determination of the survival fraction, inset shows the zoom in plot of the graph. Theminimumoccurs at ε=0.2294.

7

Nano Futures 2 (2018) 045005 MNazari et al

årq e

=-

e=

( ( ) ) ( )P x

N. 5

i

NI i

1

So after treatment, there is ρε×N number of events in the treated sample which have the probability largerthan ε to be counted as the intact viruses. This count can bewritten as

r e´ = ´ - + - ´e e( ) ( ) ( )N n N n q1 , 6

where n in the first part is the actual number of intact viruses in the treated sample, and qε is the probability ofcounting a broken virus as an intact one in the treated sample. In this calculation, since eq0 1, the rwhich

is n/N, will be bounded by re-

er .1

Byminimizing there-

e

1over the ε one can determine the upper bound for

survival fraction.However there is another point to explore. This value of r has a large error bar due to the shotnoise, so that the error in estimating n is proportional to the n . In practice for large enough epsilon therewillbe no data points satisfying the condition of e( )P xI T and the estimated value of ρεwould be zero. Thus as apractical approach, we consider a cutoff and vary epsilon from0 to themaximal value at which r ´e N 5,where 5 is an arbitrary threshold that controls how accurate thismethod is. As shown infigure 4(b) in our case,rmax can be estimated as 2.5×10−3 which is similar to the inverse of the reduction value∼103 reported usinginfectivity assays (figure 1(b)). So these data demonstrate that the proposed formulationmakes it possible todetermine the inactivation efficiency using the nanoporemethod; while in contrast to the conventionalinfectivity assays, the nanopore technique is capable of probing small damages to the viruses in linear base withhigh precision.

Generally the electrophoretic driving force and electroosmotic flow are themain effects that govern aparticle’smotion in a nanopore [43]. Thefirst one originates from the force exerted to the charged nanoparticlein an electric field and the second one is due to the viscous drag by thefluidflowing through the chargednanopore in response to the applied voltage. Combining the equations for electrophoretic and electroosmotictransport yields an effective virus velocity v in an external electric fieldE inside the pore as [44]

eh

z z= -( ) ( )vE

, 7virus pore

where ζ is zeta potential, η and ε are the viscosity and permittivity of the electrolyte. For our negatively chargedviruses attempting to translocate through the nanopore with negative surface charge, these two forces opposeeach other and the relativemagnitudes of the forces determines the direction and duration of transloca-tions [45].

In our study, electroosmotic flow is a dominant factor, and therefore, the difference inmembrane and virussurface charge plays a dominant role on capture and transport.Wewere not able to observe a statisticallysignificant difference between the zeta potential associatedwith treated and untreated virus samples. As showninfigure 3(e), the dwell time ofmost of the events in the treated sample are<90 μs, while untreated viruses havemuch longer dwell times. In addition, upon treatment, the dwell time distribution narrows substantially and thefractional current blockades are reduced, which can be attributed to amorphological change in the virus sample.

To further examine the impact of fs laser pulses on virus conformation,we employedDLS to characterizechanges in the radius ofΦX174 viruses and itsDNAcaused by laser treatment. TheDLSmeasured the averagediameter of theuntreated viruses 30.9±2.3 nm (figure 5(a)), which is in agreementwith the knownΦX174particle size [46]. TheDLS results demonstrate twouseful pieces of information:first it shows that the laser-treatedsample no longer possesses amonomodal size distribution and a secondpeak is observed at 458±56 nmwhichwill be discussed shortly. Another important piece of information can beobtained from the intersection of thecorrelation curveon the y-axis of the correlogram.This y-intercept is related to the signal-to-noise ratio of themeasured sample. As shown infigure 5(b) the correlation coefficient for the treatedΦX174viruses has a lowery-intercept of 0.4 compared to the untreated onewhich is∼0.85. The lower y-intercept in the treated sample canbeattributed to the concentration variations in this sample andpresence of 458 nmparticles. This concentrationvariation can also be explored bynanoporemeasurements. In case of the nanoporemeasurements, thewaitingtime between two successive events (δt) is inversely proportional to the concentrationof thedetectable particles.An exponentialfit to thewaiting timedata (figure 5(c)), indicates that capture rate of the viruses has decreasedbymore thanone order ofmagnitude after laser treatment at the same voltage bias (40mV). Capture rates foruntreated and treated sampleswere 52±0.87 s−1 and3.49±0.13 s−1 respectively.

Particles of the secondmode in theDLS data ofΦX174 viruses have a hydrodynamic radius that is too largeto be detectedwith our 38 nmnanopore. To investigate the second peak, we performedDLSmeasurement ofΦX174 virionDNA, the single-stranded viral DNA isolated frompurified phage by phenol extraction (NewEnglandBiolabs). Thismeasurement indicates if the second peak is a signature of freeDNA ejected from virusunder photonic exposure. TheDLSmeasured average diameter of the untreated phenol-extractedDNAwas113.5±30 nm.A part of this variationmay be attributed to∼15%of the phenol-extractedDNAmolecules notbeing in circular form. The radius of gyration,Rg and hydrodynamic radius,Rh. of freeΦX174DNApolymer can

8

Nano Futures 2 (2018) 045005 MNazari et al

be estimated as

= ( )RPL2

6, 8g

c

from the knownpersistence lengthP and the contour lengthLc [47–49]. ForΦX174DNAwith 5386nucleotidesand the persistence length∼4.6 nm forDNA [50], theRg andRh canbe estimated as 72 nmand45.6 nm,respectively [41, 51]. So the corresponding hydrodynamic diameter is approximately 91 nmwhich is consistentwith the peak inDLSmeasurement of untreatedDNA. Interestingly,DLSdata of laser exposedDNA (figure 5(d)),still contains this peakwhich argues against complete agglomeration of single-strandedDNAdue to laser exposure.Also, as the correlogramof theDNA (figure 5(e)) shows there is little difference in the timeautocorrelation functionbetween exposed andunexposedDNAsamples.Our observations indicate that the 458 nmpeak observed in laser-treatedΦX174viruses sample is thusnot from thepresence of free single-strandedDNA.

Based on our data after laser treatment there are three possiblemorphology changes that can happen toΦX174 viruses: (i) fragmentation to small pieces, (ii) perforation of viruses that leads toDNA expulsionwith asmall associated change in the virus shell, and (iii) agglomeration of viral particles (figure 6). Nanopore data canbe used to evaluate each of the possibilities. The fact that the fractional current blockade decreases aftertreatment, suggests that the particle diameter gets smaller upon laser exposure. Also if we neglect possiblechanges in the zeta potential of the viruses, the smaller viruses will translocate faster consistent with theobservation of a shorter dwell times in the treated sample. Therefore, the current blockade and the dwell timeboth point to detection of entities smaller than the original viruses, and support the hypothesis of virusperforation. At the same time, the nanopore capture rate suggests that viruses for themost part have beenfragmented to small pieces not detectable by the nanopore conductance.

Figure 5. (a)DLS size distribution obtained from (b) quasi-elastic light scattering intensity autocorrelation function or correlogramofΦX174 virus before (untreated) and after irradiation (treated)with IR&blue 35 fs laser pulses for 30 min (average laser power 2.5 W)(c) inter-event waiting time obtained fromnanoporemeasurement of the data infigure 3 at 40 mV applied bias,fitted by anexponential function. (d) Size distribution (e) correlogramofΦX174 viral DNA. Laser irradiation conditions were identical to (a), (b).

Figure 6. Schematic of the possible effect of the fs laser treatment onΦX174 virus.

9

Nano Futures 2 (2018) 045005 MNazari et al

NeitherDLS nor nanopore data do not suggest extensive virus agglomeration. Since the differentialscattering cross-section for elastic scattering in the Rayleigh regime is proportional to the 6th power of theparticle radius with given dielectric properties, DLSmay be expected to be sensitive to agglomeration or presenceof larger particle species. Additionally, significant agglomeration is expected to lead to a higher intensity atsecond diffractionmodewhich is not observed. Also in the nanopore data, significant agglomeration of virusescan block the porewhich leads to long lasting dwell times in contrast towhat is observed in our experiments. Sotaken together, our data suggest that a small fraction of viruses (7%) get perforated, while a negligible fraction(<10−3%based onRayleigh scattering assumption) get agglomerated, and the remaining largest fraction of theviruses (93%) are fragmented.

WhileDLS provides an ensemble averaged estimate of the particle size distribution, it does not provideinformation at the single particle level. Our experiments suggest nanopore single-particle conductancemeasurements to be helpful in detecting intact viruses in small-volume samples (∼0.5 μl) to characterizemorphological changes caused by laser treatment and inactivation.

4. Conclusion

The ultrashort pulsed laser technology presented here can be readily used for rapid and effective disinfection ofviruses in a label-freemanner. Applying this technology to disinfect viruses can lead to some changes in theviruses which needs to bemonitoredwith high precision. Nanopore technique allowed us to detect intact virusesandmonitor the effect of fs laser on amodel virus at the single virus level. Analysis of changes in the ionic currentthrough the nanopore provides information about size and physical properties of viruses before and after lasertreatment. To the best of our knowledge, this is the first time that nanopores have been used tomonitor changesinduced by an inactivationmethod in viruses.

Our data is a promising step towards developing a label-free detection technique that can also be used as aneffectivemethod formonitoring the survival fraction of viruses using low sample volume, high precision and fastassay time.

Acknowledgments

Grants: Kirill S Korolevwas partially supported byCottrell Scholar Award (#24010) by the ResearchCorporation for Science Advancement, a grant from the Simons Foundation (#409704) and an award fromGordon andBettyMoore foundation (#6790.08). X L andHYwere supported by theChina ScholarshipCouncil (CSC), andMAAwas supported by theNational Institutes ofHealth (R01-HG009186).

ORCID iDs

Shyamsunder Erramilli https://orcid.org/0000-0003-3950-9122

References

[1] AuBuchon J 2011Update on the status of pathogen inactivationmethods ISBT Sci. Ser. 6 181–8[2] Hellstern P 2004 Solvent/detergent-treated plasma: composition, efficacy, and safetyCurr. Opin. Hematology 11 346–50[3] RockG2011A comparison ofmethods of pathogen inactivation of FFPVox Sanguinis 100 169–78[4] NazariM, XiM, Lerch S, AlizadehM, Ettinger C, AkiyamaH,Gillespie C,Gummuluru S, Erramilli S andReinhard BM2017

Plasmonic enhancement of selective photonic virus inactivation Sci. Rep. 7 11951[5] TsenK-T, Tsen S-WD,ChangC-L,HungC-F,WuT-C andKiang JG 2007 Inactivation of viruses by laser-driven coherent excitations

via impulsive stimulated Raman scattering process J. Biomed. Opt. 12 064030–6[6] TsenKT, Tsen S-WD, FuQ, Lindsay SM, Li Z, Cope S, Vaiana S andKiang JG 2011 Studies of inactivation of encephalomyocarditis

virus,M13 bacteriophage, and Salmonella typhimuriumby using a visible femtosecond laser: insight into the possible inactivationmechanisms J. Biomed. Opt. 16 078003–8

[7] Tsen S-WD,Chapa T, BeattyW, TsenK-T, YuD andAchilefu S 2012 Inactivation of enveloped virus by laser-driven proteinaggregation J. Biomed. Opt. 17 128002

[8] Tsen S-WD,KingsleyDH, Poweleit C, Achilefu S, SorokaD S,WuT andTsenK-T 2014 Studies of inactivationmechanismof non-enveloped icosahedral virus by a visible ultrashort pulsed laserVirology J. 11 20

[9] TsenK, Tsen S-WD,HungC-F,WuT andKiang JG 2008 Selective inactivation of human immunodeficiency virus withsubpicosecond near-infrared laser pulses J. Phys.: Condens.Matter 20 252205

[10] Tsen S-WD,Tsen Y-SD, TsenK andWuT2010 Selective inactivation of viruses with femtosecond laser pulses and its potential use forin vitro therapy J. Healthc. Eng. 1 185–96

[11] KangroHOandMahy BW1996VirologyMethodsManual (NewYork: Academic)[12] DaaboulG, Yurt A, ZhangX,HwangG,Goldberg B andUnluM2010High-throughput detection and sizing of individual low-index

nanoparticles and viruses for pathogen identificationNano Lett. 10 4727–31

10

Nano Futures 2 (2018) 045005 MNazari et al

[13] Mitra A,Deutsch B, Ignatovich F,Dykes C andNovotny L 2010Nano-optofluidic detection of single viruses and nanoparticlesACSNano 4 1305–12

[14] Stoffel C L, Kathy R F andRowlenKL 2005Design and characterization of a compact dual channel virus counterCytometryA 65 140–7[15] HarmsZD,HaywoodDG, Kneller AR, Selzer L, Zlotnick A and Jacobson SC2014 Single-particle electrophoresis in nanochannels

Anal. Chem. 87 699–705[16] PiruskaA, GongM, Sweedler J V andBohnPW2010Nanofluidics in chemical analysisChem. Soc. Rev. 39 1060–72[17] Cressiot B,Oukhaled A, PatriarcheG, Pastoriza-GallegoM, Betton J-M,Auvray LC,MuthukumarM, Bacri L and Pelta J 2012 Protein

transport through a narrow solid-state nanopore at high voltage: experiments and theoryACSNano 6 6236–43[18] FreedmanK J,Haq SR, Edel J B, Jemth P andKimMJ 2013 Singlemolecule unfolding and stretching of protein domains inside a solid-

state nanopore by electricfield Sci. Rep. 3 1638[19] FreedmanK J, JürgensM, PrabhuA, AhnCW, Jemth P, Edel J B andKimMJ 2011Chemical, thermal, and electric field induced

unfolding of single proteinmolecules studied using nanoporesAnal. Chem. 83 5137–44[20] LeddenB, FologeaD, TalagaD S and Li J 2011Nanopores (Berlin: Springer) pp 129–50[21] Li J, FologeaD, Rollings R and LeddenB 2014Characterization of protein unfoldingwith solid-state nanopores Protein Peptide Lett. 21

256–65[22] OukhaledA, Cressiot B, Bacri L, Pastoriza-GallegoM, Betton J-M, Bourhis E, Jede R,Gierak J, Auvray LC andPelta J 2011Dynamics of

completely unfolded and native proteins through solid-state nanopores as a function of electric driving forceACSNano 5 3628–38[23] TalagaD S and Li J 2009 Single-molecule protein unfolding in solid state nanopores J. Am. Chem. Soc. 131 9287–97[24] AlibakhshiMA,Halman J R,Wilson J, Aksimentiev A, AfoninKA andWanunuM2017 Picomolar fingerprinting of nucleic acid

nanoparticles using solid-state nanoporesACSNano 11 9701–10[25] BrantonD,DeamerDW,Marziali A, BayleyH, Benner S A, Butler T,Di VentraM,Garaj S,Hibbs A andHuangX 2008The potential

and challenges of nanopore sequencingNat. Biotechnol. 26 1146–53[26] Carson S andWanunuM2015Challenges inDNAmotion control and sequence readout using nanopore devicesNanotechnology 26

074004[27] BrahaO,Gu L-Q, Zhou L, LuX, Cheley S andBayleyH2000 Simultaneous stochastic sensing of divalentmetal ionsNat. Biotechnol. 18

1005–7[28] Movileanu L, Cheley S andBayleyH2003 Partitioning of individualflexible polymers into a nanoscopic protein poreBiophys. J. 85

897–910[29] ZhouK, Li L, Tan Z, Zlotnick A and Jacobson SC2011Characterization of hepatitis B virus capsids by resistive-pulse sensing J. Am.

Chem. Soc. 133 1618–21[30] HarmsZD,MogensenKB,Nunes P S, ZhouK,Hildenbrand BW,Mitra I, TanZ, ZlotnickA, Kutter J P and Jacobson SC 2011

Nanofluidic devices with two pores in series for resistive-pulse sensing of single virus capsidsAnal. Chem. 83 9573–8[31] ArjmandiN,VanRoyWand Lagae L 2014Measuringmass of nanoparticles and viruses in liquids with nanometer-scale poresAnal.

Chem. 86 4637–41[32] ArjmandiN,VanRoyW, Lagae L andBorghsG 2012Measuring the electric charge and zeta potential of nanometer-sized objects using

pyramidal-shaped nanoporesAnal. Chem. 84 8490–6[33] WuH,ChenY, ZhouQ,WangR, Xia B,MaD, LuoK and LiuQ2016Translocation of rigid rod-shaped virus through various solid-

state nanoporesAnal. Chem. 88 2502–10[34] TsenK-T, Tsen S-WD, FuQ, Lindsay SM,Kibler K, Jacobs B,WuTC,KaranamB, Jagu S andRodenRB2009 Photonic approach to

the selective inactivation of viruses with a near-infrared subpicosecond fiber laser J. Biomed. Opt. 14 064042–7[35] Thompson S S andYatesMV1999 Bacteriophage inactivation at the air-water-solid interface in dynamic batch systemsAppl. Environ.

Microbiol. 65 1186–90[36] LiuX,DaiQ, Austin L, Coutts J, Knowles G, Zou J, ChenH andHuoQ2008A one-step homogeneous immunoassay for cancer

biomarker detection using gold nanoparticle probes coupledwith dynamic light scattering J. Am. Chem. Soc. 130 2780–2[37] TsenK-T, Tsen S-WD,ChangC-L,HungC-F,WuT andKiang JG 2007 Inactivation of viruses by coherent excitations with a low

power visible femtosecond laserVirology J. 4 50[38] Tsen S-WD,WuTC,Kiang JG andTsenK-T 2012 Prospects for a novel ultrashort pulsed laser technology for pathogen inactivation

J. Biomed. Sci. 19 62[39] BroughH,AntoniouC, Carter J, Jakubik J, XuY andLutzH 2002 Performance of a novel viresolveNFR virus filterBiotechnol. Prog. 18

782–95[40] KimMJ,WanunuM, Bell DC andMeller A 2006Rapid fabrication of uniformly sized nanopores and nanopore arrays for parallel

DNA analysisAdv.Mater. 18 3149–53[41] Waduge P,HuR, Bandarkar P, YamazakiH, Cressiot B, ZhaoQ,Whitford PC andWanunuM2017Nanopore-basedmeasurements

of protein size,fluctuations, and conformational changesACSNano 11 5706–16[42] Larkin J, Henley RY,MuthukumarM,Rosenstein J K andWanunuM2014High-bandwidth protein analysis using solid-state

nanoporesBiophys. J. 106 696–704[43] HuangG,WillemsK, SoskineM,WlokaC andMaglia G 2017 Electro-osmotic capture and ionic discrimination of peptide and protein

biomarkers with FraC nanoporesNat. Commun. 8 935[44] FirnkesM, PedoneD,Knezevic J, DöblingerM andRantU 2010 Electrically facilitated translocations of proteins through silicon

nitride nanopores: conjoint and competitive action of diffusion, electrophoresis, and electroosmosisNano Lett. 10 2162–21627[45] MelnikovDV,Hulings ZK andGrachevaME2017 Electro-osmotic flow through nanopores in thin andultrathinmembranes Phys.

Rev.E 95 063105[46] BayerMandDeBlois R 1974Diffusion constant and dimension of bacteriophageΦX174 as determined by self-beat laser light

spectroscopy and electronmicroscopy J. Virology 14 975–80[47] BoalD andBoalDH2012Mechanics of the Cell (Cambridge: CambridgeUniversity Press)[48] Rechendorff K,WitzG, Adamcik J andDietler G 2009 Persistence length and scaling properties of single-strandedDNA adsorbed on

modified graphite J. Chem. Phys. 131 09B604[49] SimAY, Lipfert J, HerschlagD andDoniach S 2012 Salt dependence of the radius of gyration and flexibility of single-strandedDNA in

solution probed by small-angle x-ray scatteringPhys. Rev.E 86 021901[50] ManningGS 2006The persistence length ofDNA is reached from the persistence length of its null isomer through an internal

electrostatic stretching forceBiophys. J. 91 3607–16[51] TreeDR,Muralidhar A,Doyle P S andDorfmanKD2013 IsDNA a goodmodel polymer?Macromolecules 46 8369–82

11

Nano Futures 2 (2018) 045005 MNazari et al